4 ms·
Good point. But "writes" are very fast, in our tests write latency is less than half read latency, so we can easily do master to slave replication within the SL
by damienkatz 13y ago
Good point. But "writes" are very fast, in our tests write latency is less than half read latency, so we can easily do master to slave replication within the SLA. But you point is correct, a Dynamo system is faster to achieve the same replication factor.
- jbellis 13y agoI think it's worth elaborating that the primary advantage to the dynamo model is not in the best- or average- case, but when everything does not go as planned -- when the master gets behind, when ec2 network latency spikes, etc. Then "any two replicas" instead of "master plus one more" is much more robust.
- banachtarski 13y agoDo you surround the word "writes" in quotes because they aren't truly writes in the durable sense?